Genetics issues in ancient India

[Steve Farmer]

Vidyasankar Sundaresan writes:

> On another note of caution, those who eagerly look forward
> to genetics solving the problem of Aryan origins (on both
> sides of the debate), should be prepared to be disappointed
> somewhat, with respect to time scales. It is not going to
> be easy. The science usually talks in terms of many kilo
> years (kya). As far as I can see from the current state of
> research, the conclusions of this field of science are more
> definitive about trends from 6000-10000 years ago (and more)
> as compared to events from 3000 years ago. Genetic evidence
> tends to be more reliable for periods over which evolutionary
> changes take place, and also for extremely short time periods,
> spanning 3-4 generations, but they can be sometimes quite
> inconclusive for the 5000 years of recorded human history.

I'm in full agreement with this and the rest of your last post.
This key point was missed when the paper by Kivisild et al.
came out last year (Current Biology 8 Nov. 1999,
9[22]:1331-1334), which focused on early migrational patterns
into India. When the paper first appeared, it
was widely cited as providing evidence against any Aryan
migration (e.g., by Disotell TR, Current Biology 9[24]: R925-8).
What no one bothered to note was that the divergence time that
Kivisild found for the most recent influx of the putative
Caucasoid haplogroup U into India (putative since this too is in
dispute) was given within errors of + or - 3,000 years BP.

Error ranges of up to 6,000 years, needless to say, can wipe
out a big hunk of history.

When you systematically review these studies, you get the
impression that linguistics may remain the best evidence for
studying the Aryan migration issue for quite some time. Another
common error in these papers -- one shared by the Kivisild &
Disotell studies -- is to assume close overlap between genes and
language, ignoring acculturation issues. On this obvious but
often overlooked point, I cite a post from Mark Hall this
morning on the NEAsianStudies List:

> Since it has come up repeatedly on this list and others, the equating of
> genes, pots, peoples and languages, an interesting article that hasn't seen
> much press is:
>
> Comas, Calafell, et al. (2000) Georgian and Kurd mtDNA sequence analysis
> shows a lack of correlation between languages and female genetic lineages,
> AMERICAN JOURNAL OF PHYSICAL ANTHROPOLOGY Vol. 112, 5-16.
>
> In a quick summary, using 360 base pairs from the mtDNA sequence, they
> looked at the mtDNA from 45 Georgians and 29 Kurds.  Georgians are
> Kartvelian speakers while the Kurds are Indo-Iranian speakers. The genetic
> analysis points only to minor differences between the Georgians and Kurds
> and the rest of the European populations and very different from the Basque
> population (which Georgians are often linked to).  Their conclusion is the
> linking of European mtDNA with the Indo-European languages is seriously
> questioned.

Hopefully, enough on this topic for now. The potentials for abuse of
these studies in studying migrational issues are manifold -- and
not
anything suitable for detailed discussion on this List.

Steve Farmer